We examined the relationships between folate and methionine intake, serum homocysteine levels (as a biomarker for folate metabolism), and methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism genotype and risk of oral cancer in a population-based, case-control study in Puerto Rico. Structured questionnaires were used to collect information on demographic factors, usual adult diet, and tobacco and alcohol use. Oral epithelial cells and blood samples were collected from a subset of subjects. Analyses were conducted by logistic regression, adjusting for age, sex, lifetime smoking and lifetime alcohol intake, with the following numbers of cases/controls, respectively: dietary data (341/521); MTHFR genotype (148/149); and homocysteine (60/90). Although increased folate intake was associated with decreased oral cancer risk [adjusted odds ratio (OR) in highest vs. lowest quartile = 0.6, 95% confidence interval (CI): 0.4, 1.0, P(trend) = 0.05)], this finding was due almost entirely to folate intake from fruit (adjusted OR = 0.4, 95% CI: 0.2, 0.6; P(trend) = 0.0001), whereas other dietary folate sources showed no clear association. Methionine intake and serum homocysteine levels were not associated with oral cancer risk. Subjects with the MTHFR C677T homozygous variant (TT) genotype had a nonsignificantly lower risk, and risk patterns tended to differ by level of folate, methionine, alcohol intake and smoking, although the power to detect significant associations in subgroups of these variables was low. Risks for oral cancer are not folate specific; preventive recommendations for this disease should emphasize the importance of a healthy diet, including substantial intake of fruits.